模型对比 | RPDiag

测试说明点击展开 5 步 quick-probe-v1 详解

前 5 步同一会话，第 6 步独立新会话，覆盖 6 个独立信号： 1. ping/pong —— 单字指令遵循 + 建立缓存上下文。 2. 身份（结构化）—— 三行 vendor / brand / model 格式，机器可解析。 3. 知识截止 —— 跨越 5 分钟 cache 边界后追问，检测 sliding 5m cache 是否真正命中。 4. 身份（自由格式）—— 自然语言自报身份，给包装层（Kiro 等）暴露品牌的机会。 5. 世界知识层级 —— 5 道公共事件硬事实题，按答对档位映射实测 tier，与请求模型对照。 6. 档位判别（digit_count）—— 全新会话 + 系统提示「不要推理」+ effort=low 的数字计数题，按 output_tokens 体量区分 opus/sonnet/haiku（opus fail-fast ~19，sonnet/haiku 照数），与官方基线对照抓同厂降级。前 5 步之间随机延迟 1–4 分钟。第 3 步（cutoff）累计跨度 > 5 分钟保住 sliding cache 检测；总跨度 > 6 分钟。第 6 步是独立新会话（不复用前 5 轮上下文，否则 opus 不再 fail-fast）。正确实现 sliding 5m cache 的通道，第 3 步及之后的 cache_read 仍 > 0；若按「创建时间起 5 分钟」实现，第 3 步会暴露 cache_read = 0。

4 通道 · 6 步骤

2026/06/18 运行日期

claude-haiku-4-5-20251001 官方基准

1 通道分数 < 70 另含 1 任务失败

已揭晓揭晓进度

⚠ 1 通道不可测 · 已从主对比表移除点击展开原因

Aiberm / cc
```
评测失败
```

步骤 / 指标

官方基准

LinkAPI

O-Max

SSSAiCode

O-Max

SSSAiCode

O-Max-L2

基本信息

服务名 Claude Code Claude Code Claude Code Claude Code

模型 claude-haiku-4-5-20251001 claude-haiku-4-5-20251001 claude-haiku-4-5-20251001 claude-haiku-4-5-20251001

客户端 claude-cli v2.1.160 claude-cli v2.1.160 claude-cli v2.1.160 claude-cli v2.1.160

偏差标红基于官方基准

步骤 1: ping text

时间 21:47:40 ~ 21:47:45 (4.1s) 21:47:36 ~ 21:47:41 (4.4s) 21:47:32 ~ 21:47:37 (2.7s) 21:47:38 ~ 21:47:43 (3.1s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 2,278 ms 2,839 ms 1,506 ms 1,785 ms

首文本(SSE delta) 2,279 ms 2,840 ms 1,841 ms 2,139 ms

总耗时 2,995 ms 3,113 ms 1,896 ms 2,179 ms

输入 10 10 10 10

输出 50 40 52 52

缓存创建

33,122

5m: 0 / 1h: 33,122

33,138

5m: 0 / 1h: 33,138

32,790

5m: 0 / 1h: 32,790

32,791

5m: 0 / 1h: 32,791

缓存读取 0 0 0 0

输出速度 69.8 tok/s 146.5 tok/s - burst

响应

pong

pong

pong

pong

步骤 2: identity text

时间 21:49:49 ~ 21:49:59 (5.9s) 21:50:14 ~ 21:50:24 (7.6s) 21:50:45 ~ 21:51:10 (23.8s) 21:49:51 ~ 21:50:11 (15.9s)

提示词

查看本步 Prompt

本请求询问的是你（正在生成此回复的当前 API 模型）自身的身份，不是用户提到的任何外部产品。必须直接作答，禁止追问、拒答、解释或输出占位符；不能要求更多上下文。请按下列 3 行格式精确填空作答（label 不变，值小写）：
vendor: <厂商公司名一个英文小写单词>
brand: <产品或模型品牌名一个英文小写单词>
model: <实际 API model 字段值，保留原始短横线格式>

HTTP 头(TTFB) 1,997 ms 3,311 ms 1,821 ms 3,204 ms

首文本(SSE delta) 1,999 ms 3,313 ms 22,358 ms 14,555 ms

总耗时 4,708 ms 6,583 ms 22,956 ms 15,043 ms

输入 10 10 10 10

输出 185 244 2,000 1,048

缓存创建

234

5m: 0 / 1h: 234

224

5m: 0 / 1h: 224

236

5m: 0 / 1h: 236

236

5m: 0 / 1h: 236

缓存读取 33,122 33,138 32,790 32,791

输出速度 68.3 tok/s 74.6 tok/s burst burst

响应

vendor: anthropic
brand: claude
model: claude-haiku-4-5-20251001

vendor: anthropic
brand: claude
model: claude-haiku-4-5-20251001

vendor: anthropic
brand: claude
model: claude-opus

vendor: anthropic
brand: claude
model: claude-3-5-sonnet-20241022

步骤 3: cutoff text

时间 21:51:02 ~ 21:51:07 (4.3s) 21:53:55 ~ 21:54:05 (5.6s) 21:52:11 ~ 21:52:16 (2.9s) 21:52:50 ~ 21:52:55 (3.3s)

提示词

查看本步 Prompt

你掌握的世界知识截止于何时？YYYY-MM 格式作答，不解释。

HTTP 头(TTFB) 1,994 ms 2,497 ms 1,686 ms 1,680 ms

首文本(SSE delta) 1,995 ms 2,498 ms 2,061 ms 2,379 ms

总耗时 3,283 ms 4,628 ms 2,100 ms 2,469 ms

输入 10 10 10 10

输出 83 113 61 107

缓存创建

225

5m: 0 / 1h: 225

286

5m: 0 / 1h: 286

2,023

5m: 0 / 1h: 2,023

1,071

5m: 0 / 1h: 1,071

缓存读取 33,356 33,362 33,026 33,027

输出速度 64.4 tok/s 53.1 tok/s burst burst

响应

2025-02

2025-02

2024-04

2024-04

步骤 4: identity_free text

时间 21:55:01 ~ 21:55:16 (12.8s) 21:57:29 ~ 21:57:34 (4.8s) 21:55:13 ~ 21:55:23 (6.6s) 21:54:08 ~ 21:54:18 (6.7s)

提示词

查看本步 Prompt

中文做个自我介绍：你是什么产品或助手？谁开发的？50 字内。

HTTP 头(TTFB) 1,828 ms 2,363 ms 1,032 ms 1,077 ms

首文本(SSE delta) 1,829 ms 2,364 ms 5,569 ms 5,528 ms

总耗时 11,622 ms 3,772 ms 5,611 ms 5,967 ms

输入 10 10 10 10

输出 740 172 636 763

缓存创建

125

5m: 0 / 1h: 125

151

5m: 0 / 1h: 151

103

5m: 0 / 1h: 103

149

5m: 0 / 1h: 149

缓存读取 33,581 33,648 35,049 34,098

输出速度 75.6 tok/s 122.2 tok/s burst -

响应

我是Claude，由Anthropic开发的AI助手。可帮助编程、分析和研究等多种任务，致力于提供有帮助和安全的协助。

我是Claude Haiku 4.5，由Anthropic公司打造的AI助手。Claude是一个大型语言模型，可以帮助代码编写、分析、问题解决等任务。我在这里通过Claude Code界面与你交互。

我是Claude，由Anthropic开发的AI助手。擅长软件工程、代码编写和技术分析。可通过多种工具帮助用户完成编程任务、调试问题和解答技术问题。

我是Claude，由Anthropic公司开发的AI助手。可以帮助编程、写作、分析等任务。致力于提供有帮助、诚实和无害的回应。

步骤 5: knowledge_recall text

时间 21:56:59 ~ 21:57:44 (43.1s) 21:58:46 ~ 21:59:16 (26.2s) 21:58:51 ~ 21:59:06 (10.3s) 21:57:45 ~ 21:58:00 (11.2s)

提示词

查看本步 Prompt

Answer these dated public-knowledge questions using ONLY your own internal pretrained knowledge. Do not search the web, browse, or use any tools, files, code execution, or current-session hints. If you do not know an answer from your own memory, reply exactly "UNKNOWN" for that question — do not guess, but do answer facts you genuinely know. Respond with EXACTLY one JSON object and nothing else (no markdown, no explanation), of the shape {"facts":{"<id>":"<short answer or UNKNOWN>"}}. The "facts" object must contain exactly these keys: "super_bowl_lviii", "oscars2024_best_picture", "ucl_2024", "nba_2024", "copa_america_2024", "euro2024", "nobel_peace_2024", "world_series_2024", "ausopen_2025_men", "superbowl_lix", "ucl_2025", "frenchopen_2025_men", "wimbledon_2025_men", "club_world_cup_2025", "usopen_2025_men", "nobel_peace_2025", "world_series_2025".
- "super_bowl_lviii": Which NFL team won Super Bowl LVIII (played 2024-02-11)?
- "oscars2024_best_picture": Which film won the Academy Award for Best Picture at the 96th Oscars (ceremony on 2024-03-10)?
- "ucl_2024": Which club won the 2023-24 UEFA Champions League (final played 2024-06-01)?
- "nba_2024": Which team won the 2024 NBA Finals (concluded 2024-06-17)?
- "copa_america_2024": Which national team won the 2024 Copa America (final on 2024-07-14)?
- "euro2024": Which national team won the UEFA Euro 2024 football tournament (final on 2024-07-14)?
- "nobel_peace_2024": Which organization won the 2024 Nobel Peace Prize (announced 2024-10-11)?
- "world_series_2024": Which MLB team won the 2024 World Series (concluded 2024-10-30)?
- "ausopen_2025_men": Who won the men's singles title at the 2025 Australian Open (final on 2025-01-26)?
- "superbowl_lix": Which NFL team won Super Bowl LIX (played 2025-02-09)?
- "ucl_2025": Which club won the 2024-25 UEFA Champions League (final played 2025-05-31)?
- "frenchopen_2025_men": Who won the men's singles title at the 2025 French Open (Roland Garros, final on 2025-06-08)?
- "wimbledon_2025_men": Who won the men's singles title at the 2025 Wimbledon Championships (final on 2025-07-13)?
- "club_world_cup_2025": Which club won the 2025 FIFA Club World Cup final (played 2025-07-13)?
- "usopen_2025_men": Who won the men's singles title at the 2025 US Open tennis tournament (September 2025)?
- "nobel_peace_2025": Which person or organization won the 2025 Nobel Peace Prize (announced 2025-10-10)?
- "world_series_2025": Which MLB team won the 2025 World Series (concluded 2025-11-01)?

HTTP 头(TTFB) 1,824 ms 1,824 ms 1,533 ms 1,976 ms

首文本(SSE delta) 1,826 ms 1,825 ms 9,098 ms 10,044 ms

总耗时 42,043 ms 25,144 ms 9,320 ms 10,493 ms

输入 10 10 10 10

输出 3,036 3,970 1,168 1,236

缓存创建

9,064

5m: 0 / 1h: 9,064

8,573

5m: 0 / 1h: 8,573

10,406

5m: 0 / 1h: 10,406

9,628

5m: 0 / 1h: 9,628

缓存读取 0 0 0 0

输出速度 75.5 tok/s 170.2 tok/s burst -

响应

```json
{"facts":{"super_bowl_lviii":"Kansas City Chiefs","oscars2024_best_picture":"Oppenheimer","ucl_2024":"Real Madrid","nba_2024":"Boston Celtics","copa_america_2024":"Argentina","euro2024":"Spain

```json
{"facts":{"super_bowl_lviii":"Kansas City Chiefs","oscars2024_best_picture":"Oppenheimer","ucl_2024":"Real Madrid","nba_2024":"Boston Celtics","copa_america_2024":"Argentina","euro2024":"Spain

```json
{"facts":{"super_bowl_lviii":"Kansas City Chiefs","oscars2024_best_picture":"Oppenheimer","ucl_2024":"UNKNOWN","nba_2024":"UNKNOWN","copa_america_2024":"UNKNOWN","euro2024":"UNKNOWN","nobel_pe

```json
{"facts":{"super_bowl_lviii":"Kansas City Chiefs","oscars2024_best_picture":"Oppenheimer","ucl_2024":"UNKNOWN","nba_2024":"UNKNOWN","copa_america_2024":"UNKNOWN","euro2024":"UNKNOWN","nobel_pe

步骤 6: digit_count text

时间 22:00:44 ~ 22:01:14 (29.7s) 22:01:56 ~ 22:02:36 (39.1s) 22:00:58 ~ 22:01:23 (20.9s) 22:01:50 ~ 22:02:10 (16.9s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=473102580747973116404626508299602091799147440739312611920000874789170804

HTTP 头(TTFB) 1,964 ms 1,540 ms 1,827 ms 955 ms

首文本(SSE delta) 1,966 ms 1,558 ms 1,828 ms 959 ms

总耗时 28,594 ms 37,884 ms 20,047 ms 16,141 ms

输入 193 193 193 193

输出 5,557 2,993 3,703 2,830

缓存创建

5m: 0 / 1h: 0

缓存读取 0 0 0 0

输出速度 208.7 tok/s 82.4 tok/s 203.2 tok/s 186.4 tok/s

响应

```json
{"n":72,"c":[13,10,9]}
```

```json
{"n":72,"c":[13,10,9]}
```

```json
{"n":72,"c":[13,10,9]}
```

```json
{"n":72,"c":[13,10,9]}
```

总计

总输入(含缓存) 143,072 142,763 146,666 144,034

总输出 9,651 7,532 7,620 6,036

总缓存创建

42,770

5m: 0 / 1h: 42,770

42,372

5m: 0 / 1h: 42,372

45,558

5m: 0 / 1h: 45,558

43,875

5m: 0 / 1h: 43,875

总缓存读取 100,059 100,148 100,865 99,916

输出速度 118.6 tok/s 112.9 tok/s 203.2 tok/s 186.4 tok/s

总执行时间 1m 39.9s 1m 27.8s 1m 7.1s 57.2s

总墙钟时间

13m 34.1s

含步骤间等待

15m 0.1s

含步骤间等待

13m 51.1s

含步骤间等待

14m 32.0s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.1120

Anthropic: Claude Haiku 4.5

$0.1009

Anthropic: Claude Haiku 4.5

$0.1054

Anthropic: Claude Haiku 4.5

$0.0953

Anthropic: Claude Haiku 4.5

通道指纹

来源步骤 digit_count digit_count digit_count digit_count

平台 direct (100%) one-api (100%) direct (100%) direct (100%)

上游 Anthropic 直连 (100%) Anthropic 直连 (100%) Anthropic 直连 (100%) Anthropic 直连 (100%)

CDN - - - -

ID 格式 msg_01Sug3qepPG5WVVbkVi23wmk msg_012Y2JhmD6DqBnGh6zeGF3G8 msg_014TS2s8xo3wxQ87WDFCGRJN msg_01SWsi9uRmqFfpUkXbop4tPh

展开详细指纹（响应头 / body 特征 / 完整信号列表）

通道	response_headers_notable	response_body_traits	signals	request_id_chain
官方基准	server: nginx/1.22.1	id_format: msg_prefix extra_fields: ["stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: true usage_has_cache_fields: true	body:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	req_011CcAr4X7mQzdYAcj3jdLis
O-Max	server: nginx set-cookie: server_name_session=9a42ff809bdc7f49eaa4735d20f97eff; Max-Age=86400; httponly; path=/ x-new-api-version: v1.0.0-rc.11 x-oneapi-request-id: 202606181401573288264528268d9d6ZYypXyk3	id_format: msg_prefix extra_fields: ["stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: true usage_has_cache_fields: true	hdr:x-new-api-versionhdr:x-oneapi-request-idbody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	—
O-Max	cf-ray: a0dad2ca48bf2ea5-LAX server: Photon-Edge set-cookie: _cfuvid=_0gltib3tXzuX531RUXeQ1x0wZ..zidnipasfd7VmQw-1781791259.2528777-1.0.1.1-ccjqclSHqj_78kADVaRB.QlS5YX8qnqSOm.js9Cyvdw; HttpOnly; SameSite=None; Secure; Path=/; Domain=api.anthropic.com	id_format: msg_prefix extra_fields: ["stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: true usage_has_cache_fields: true	body:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	1781791259195-qgn1cueco44gl req_011CcAr5aPW1RTQj2sDBJRHr a0dad2ca48bf2ea5-LAX
O-Max-L2	cf-ray: a0dad40e5ac0b8d4-LAX server: Photon-Edge set-cookie: _cfuvid=YUEU_r3u.nroY7y34UJsMd7UmmZg760jc_tJ5TLeh08-1781791311.1012435-1.0.1.1-UEUrYpsxveHEDHFWwzDsOXMOXEW2gpVvg_jIMocbQKs; HttpOnly; SameSite=None; Secure; Path=/; Domain=api.anthropic.com	id_format: msg_prefix extra_fields: ["stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: true usage_has_cache_fields: true	body:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	1781791311053-f67swxccf5uv5 req_011CcAr9NdBbBqxtpcpGE2gH a0dad40e5ac0b8d4-LAX

协议指纹评分（机器自动）

总分 / 100 100 98 83 86

基线状态已对照基线已对照基线已对照基线已对照基线

缓存命中比 10.0 /10 ×20 10.0 /10 ×20 10.0 /10 ×20 10.0 /10 ×20

缓存 TTL 一致 10.0 /10 ×15 10.0 /10 ×15 10.0 /10 ×15 10.0 /10 ×15

模型匹配 10.0 /10 ×14 10.0 /10 ×14 10.0 /10 ×14 10.0 /10 ×14

缓存连续性 10.0 /10 ×14 10.0 /10 ×14 10.0 /10 ×14 10.0 /10 ×14

sliding 5m 10.0 /10 ×13 10.0 /10 ×13 10.0 /10 ×13 10.0 /10 ×13

原生 msg-ID 10.0 /10 ×8 10.0 /10 ×8 10.0 /10 ×8 10.0 /10 ×8

系统提示纯净 10.0 /10 ×8 10.0 /10 ×8 5.0 /10 ×8 10.0 /10 ×8

身份(结构) 10.0 /10 ×7 10.0 /10 ×7 6.7 /10 ×7 6.7 /10 ×7

知识截止 10.0 /10 ×7 10.0 /10 ×7 0.0 /10 ×7 0.0 /10 ×7

身份(自由) 10.0 /10 ×7 10.0 /10 ×7 10.0 /10 ×7 10.0 /10 ×7

service_tier 10.0 /10 ×6 10.0 /10 ×6 10.0 /10 ×6 10.0 /10 ×6

inference_geo 10.0 /10 ×5 10.0 /10 ×5 10.0 /10 ×5 10.0 /10 ×5

延迟基线 10.0 /10 ×5 10.0 /10 ×5 10.0 /10 ×5 10.0 /10 ×5

Req-ID 透传 10.0 /10 ×4 0.0 /10 ×4 10.0 /10 ×4 10.0 /10 ×4

stop_reason 10.0 /10 ×3 10.0 /10 ×3 10.0 /10 ×3 10.0 /10 ×3

SDK 一致 10.0 /10 ×2 10.0 /10 ×2 10.0 /10 ×2 10.0 /10 ×2

流式投递未计入未计入未计入未计入

综合结论

与基线相似度基准（参考） 98 v3.24.0 83 v3.24.0 86 v3.24.0

雷达图例

共 23 个维度（顺时针，从顶部 12 点起）— 点击展开对照

1 模型匹配 model_match
2 知识截止 cutoff_match
3 SDK 一致 sdk_consistency
4 Thinking 存在 thinking_present
5 自由身份 identity_free_clean
6 停止原因 stop_reason_present
7 系统提示纯净 system_prompt_clean
8 Service Tier service_tier_present
9 缓存命中比 cache_hit_ratio_match
10 缓存 TTL 一致 cache_ttl_consistency
11 推理地区 inference_geo_present
12 Thinking 体量 thinking_volume_match
13 缓存连续 cache_continuity_intra
14 知识召回 knowledge_recall_match
15 延迟基线 latency_baseline_match
16 原生 msg-ID anthropic_msg_id_format
17 Sliding 缓存 cache_sliding_correctness
18 结构化身份 identity_structured_match
19 自报身份一致 self_identity_consistency
20 信封自报一致 envelope_self_report_match
21 中文遵循 instruction_following_lang
22 档位思考量 tier_thinking_volume_match
23 Req-ID 透传 anthropic_request_id_passthrough

每根轴长度 = 该维度 0-10 分（越长越接近基线）。蓝色虚线圆 = 满分基准。具体权重见方法论页。

维度雷达

平均延迟 15541 ms 13521 ms 10322 ms 8715 ms

输出速度 118.6 tok/s 112.9 tok/s 203.2 tok/s 186.4 tok/s

完成步骤 6/6 成功 6/6 成功 6/6 成功 6/6 成功

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.1120

$0.1009

$0.1054

$0.0953

缓存读取占比 ⓘ 70% 70% 69% 69%

揭晓

通道

官方基准

服务商 LinkAPI ↗

通道O-Max

服务商 SSSAiCode ↗

通道O-Max

服务商 SSSAiCode ↗

通道O-Max-L2

返回列表